11:00 〜 11:15
[AOS15-02] Impact of large distributed sensor networks on global wave forecasts
★Invited Papers
キーワード：Spotter buoy, wave forecasting, global sensor network
In-situ open ocean wave observations have historically been exceedingly sparse. Consequently, despite the development of advanced data assimilation strategies, data availability limits the potential for model accuracy improvements, with major consequences for coastal communities and maritime industries. Starting in 2019, Sofar has been deploying a global network of surface drifters to increase data density in the world's oceans. Each node in this network consists of a Spotter buoy: a basketball-sized, solar-powered buoy that provides hourly observations of wave spectra (including directional moments), as well as sea surface temperature, barometric pressure, drift and surface winds. Since its inception, this network has grown to more than 700 units worldwide, and is expected to grow to 1500 units by the end of the year.
In this work we present the network, its performance and expansion, and specifically its application to wave forecasting. Sofar runs a global WaveWatch3 model that uses sequential data assimilation of wave observations to improve forecast accuracy. Recently (January 2022), and unique to our forecast, this system has switched to using spectral data from the buoys, rather than wave height only, yielding greatly improved forecast skill. A one-month-long re-analysis comparing the baseline non-assimilative model, significant wave height-based assimilation, and the novel assimilation on a spectral per-frequency basis illustrates improvements in bulk parameter predictions up to four-day lead times. The shift from scarce, coastal-focused spectral observations and satellite observations limited to significant wave height to global, high density coverage of full-spectra observations vastly expands the potential for improving marine weather forecasting accuracy if effective methods for utilization of this data can be developed and operationalized. In this work, we present a first step toward immediate utilization of this global sensor network for improved nowcast and forecast predictions.
In this work we present the network, its performance and expansion, and specifically its application to wave forecasting. Sofar runs a global WaveWatch3 model that uses sequential data assimilation of wave observations to improve forecast accuracy. Recently (January 2022), and unique to our forecast, this system has switched to using spectral data from the buoys, rather than wave height only, yielding greatly improved forecast skill. A one-month-long re-analysis comparing the baseline non-assimilative model, significant wave height-based assimilation, and the novel assimilation on a spectral per-frequency basis illustrates improvements in bulk parameter predictions up to four-day lead times. The shift from scarce, coastal-focused spectral observations and satellite observations limited to significant wave height to global, high density coverage of full-spectra observations vastly expands the potential for improving marine weather forecasting accuracy if effective methods for utilization of this data can be developed and operationalized. In this work, we present a first step toward immediate utilization of this global sensor network for improved nowcast and forecast predictions.